One third of the world's population is infected with Tuberculosis (TB). This large latent TB reservoir fuels a significant portion of TB disease and death, and TB disease is the leading cause of death among HIV- infected people.1 A significant portion of this latent reservoir is established in childhood; estimates from 2010 suggest that 53 million children in the 22 highest TB burden countries are infected with TB.2 Despite the magnitude of this epidemic in childhood, childhood TB is largely neglected. There are significant knowledge gaps in our understanding of the drivers and dynamics of childhood TB infection, especially among school-age children. To address these gaps this K23 proposal will use rigorous epidemiologic methods to assess individual, environmental, and structural predictors of TB infection in children to identify when, where, and why school-age children are infected with TB in rural Uganda. The following research aims have been designed to align with the didactic training and mentoring proposed in this application. In this K23 we propose to use a population-based cohort and case-control study design to address three aims: (1) to characterize the incidence and predictors of TB infection among a population-based sample of school-age children, (2) to determine the impact of perinatal HIV exposure on TB infection among HIV- uninfected school-age children, and (3) to explore the relationship between incident TB infection and childhood social network characteristics. To achieve these aims we will leverage the research infrastructure of a large NIH-funded population-based cohort of 320,000 persons residing in 32 communities in rural Uganda and Kenya, which collects detailed data on household demographics, socioeconomics, HIV status, and adult social networks. We will conduct this K23 study in 10 communities in Eastern Uganda, and it will include approximately 4,000 children ages 5-13 years. We will first measure the age-specific prevalence of TB infection with a baseline tuberculin skin test (TST) survey. Incident TB-infection will be measured in TST- negative children over three years. To assess predictors of TB-infection we will obtain measurements of household and school ventilation, indoor contact history, and social network analysis for a case-control study. The career development aims for this K23 are (1) to gain experience in the design and conduct of population-based longitudinal studies, (2) to build skills in advanced epidemiologic methods and (3) to learn methods in social network analysis and apply them to characterizing TB-transmission in childhood. This plan will be carried out with coursework and the guidance of an international team of mentors with expertise in population-based study design, clinical research, data analysis, advanced methods in epidemiology and biostatistics, and social network analysis. The research findings and career development afforded through this K23 will led to an R01 application on testing targeted interventions to decrease the latent TB-reservoir in childhood.